Digit recognition in noisy environments via a sequential GMM/SVM system

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

Volumn 1, Issue , 2002, Pages 49-52

  Fine, Shai ^a   Saon, George ^a   Gopinath, Ramesh A ^a  

^a IBM T J WATSON RESEARCH CENTER   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

BASELINE RECOGNITION SYSTEM; DIGIT RECOGNITION; ERROR CORRECTING OUTPUT CODES; GAUSSIAN MIXTURE MODEL; LINEAR DISCRIMINANT ANALYSIS; MAXIMUM LIKELIHOOD LINEAR TRANSFORM; STRUCTURAL RISK MINIMIZATION; SUPPORT VECTOR MACHINE; VITERBI SEARCH METHOD;

ERROR CORRECTION; LEARNING ALGORITHMS; LEARNING SYSTEMS; MATHEMATICAL MODELS; MAXIMUM LIKELIHOOD ESTIMATION; NEURAL NETWORKS; PROBABILITY DISTRIBUTIONS; RISK ASSESSMENT; SPEECH ANALYSIS; SPEECH CODING; SPEECH PROCESSING; SPURIOUS SIGNAL NOISE;

SPEECH RECOGNITION;

EID: 0036291381     PISSN: 15206149     EISSN: None     Source Type: Journal    
DOI: 10.1109/ICASSP.2002.5743651     Document Type: Article

References (8)

1. A hubrid gmm/svm approach to speaker identification
2. Enhancing gmm scores usng svm "hints"
3. Robust digit recognition in noisy environments: The ibm aurora 2 system
4. Solving multiclass learning problems via error-correcting output codes
5. Another approach to polychotomous classification
6. The aurora experimental framework for the performance evaluation of speech recognition systems under noisy conditions